Reproduce “Generalization Properties of Optimal Transport GANs with Latent Distribution Learning” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Giulia Luise, Massimiliano Pontil, Carlo Ciliberto

The authors study joint learning of a latent distribution and generator for Sinkhorn GANs, updating both together. In a synthetic helix task, the joint method yields a Sinkhorn gap proxy of 0.0003, versus 0.0311 with a fixed Gaussian prior. This comparison is limited to that setting; the theory bounds the rate only under exact-representation and regularity assumptions.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.